Inapt gratitude: against expansionist views

Psychologists and philosophers have written much about gratitude recently. Many of these contributions have endorsed expansionist views of gratitude, counseling agents to feel and express gratitude in many circumstances. I argue that the essential features of the moral norm of gratitude are that a beneficiary acknowledges and appreciates benefits provided by another who is acting from beneficence, and is disposed to provide a comparable benefit to the benefactor if a suitable occasion arises. The best-known philosophical version of expansionist views claims that gratitude is apt even in cases where the “benefactor” not only did not intend to benefit the other, but intended to harm her. In the psychological literature, expansionists typically do distinguish between being grateful to and being grateful that. But they also write as if there is one general character trait of gratefulness. In this paper I argue that the philosophical position considered is mistaken on conceptual and moral grounds, and that the dominant view among psychologists fails to recognize the difference between two different traits of gratitude, one a moral virtue and the other a prudential virtue.

which people are advised to be grateful. In the philosophical literature, gratitude is principally understood as a moral norm (an obligation and/or a virtue). This paper argues that there is a downside to this enthusiasm. The problem is not that there is too much gratitude, but rather finding gratitude everywhere obscures its real moral value, unless careful distinctions are maintained.

Preliminaries
There is a basic distinction between two types of gratitude: grateful to and grateful that (Walker 1981;Card 1988). The former may be designated as P1 is grateful to P2 for benefit B; the latter, as P3 is grateful for state of affairs X. When one person is grateful to another, this has been called 'benefit-triggered', 'targeted', or 'triadic' gratitude; being grateful for a state of affairs has been labeled 'generalized', 'propositional', or 'dyadic' gratitude (Lambert et al. 2009;McAleer 2012;Gulliford et al. 2013;Manela 2016a) (In this essay, I will use the terms 'targeted ' and 'propositional'.). Examples of targeted gratitude include that I am grateful to Jane for giving me a ride home from work, that I am grateful to Joan for comforting me when I was depressed, and I am grateful to Jessica for loaning me money in a time of great need. Cases of propositional gratitude include that I am grateful that the weather was perfect for our family picnic, I am grateful that I secured a job teaching philosophy when such positions were scarce, and I am grateful that my mother's cancer is in remission.
Many philosophers agree that gratitude is a moral requirement (e.g., Smith 1982Smith [1790, 68; Sidgwick 1981Sidgwick [1907Berger 1975, 300;Card 1988, 117;McConnell 1993, Chapter I;Manela 2016a, 283), and as such this refers to targeted gratitude: the beneficiary has some sort of obligation regarding the benefactor. The beneficiary's obligations are triggered by various factors: the benefactor provided (or attempted to provide) a benefit, acted from benevolence or goodwill (Sidgwick 1981(Sidgwick [1907Berger 1975, 299;Jecker 1989, 74), in so acting demonstrated moral merit (Smith 1982(Smith [1790, [74][75]78), provided the benefit freely (no strings attached), and the benefit was accepted by the beneficiary (McConnell 1993, Chapter I). While philosophers may disagree about some of the contours of the beneficiary's obligation to the benefactor, they at least include acknowledging and appreciating the moral significance of the benefactor's act, and being prepared to provide the benefactor with a comparable benefit if a suitable occasion arises (McConnell 1993, Chapter II). P1's experience of gratitude toward P2 is apt (or fitting) only if P2 provided P1 with a benefit with no strings attached, acting from benevolence and in a morally meritorious way, and P1 accepted the benefit. Apt gratitude is more than praise, since a third party could reasonably praise P2's actions (Card 1988, 119). The original beneficiary, P1, has a special reason to acknowledge and appreciate P2's conduct, and to desire to reciprocate if the opportunity arises. Apt gratitude includes both recognizing when gratitude is appropriate and recognizing what gratitude requires. Put another way, "gratitude would only count as appropriate when felt 'at the right times, about the right things, towards the right people, for the right end and in the right way'" (Morgan et al. 2015, 101, citing Aristotle 1985, 44 [1106b).

Simple cases of inapt gratitude
It is by now a well-documented phenomenon that sometimes when persons are kidnapped or taken hostage by others, the victims eventually identify with their captors, sometimes adopt their captors' views, and sometimes feel gratitude to them. This is now referred to as the Stockholm syndrome. John Stuart Mill has a remarkably prescient discussion of this phenomenon (Mill 1988(Mill [1869, 36; see also, McConnell 2017, 290-291). The gratitude in such cases is often linked to what the victims perceive as benefits they are being given, such as food and shelter. Such gratitude is clearly inapt, and the reason is because what the victims are being provided is not a benefit, or if it is it is not being provided by meritorious conduct. Some might think that the victim's gratitude exhibits the vice of servility, but such a judgment seems harsh given that the voluntariness of the victim's behavior is severely compromised.
A second simple example of inapt gratitude occurs in the gospel of Luke (18: 9-14). In a self-congratulatory fashion, the Pharisee gives thanks to God that he is not a sinner like other people. The tax collector, by contrast, asked for forgiveness for his sins. The lesson drawn from the parable is that those who exalt themselves will be humbled and those who are humble will be exalted. Robert Emmons calls this "a distorted form of gratitude" (Emmons 2007. 150). We see the same attitude in Donald Trump. When a reporter asked him during the Thanksgiving season what he was most thankful for, his response in part was "having made a tremendous difference in the country" (by which he means a positive difference). We might call this narcissistic gratitude. It is inapt because there is no benefit for which to be grateful. The Pharisee is not, we may presume, without sin, and it is highly doubtful that Mr. Trump has made a tremendous positive difference. It is also inapt because even if the Pharisee were without sin, that is not a benefit that would have been provided by someone else (on the assumption that whether an individual does or does not sin is a matter of free choice). And even if Mr. Trump had made a tremendous positive difference, being grateful to oneself makes little sense.
A third example of inapt gratitude is when a benefactor provides another person a benefit, but acts unethically in doing so. Robert Roberts describes such a case.
Let us say that I badly need a kidney transplant, and my doctor, who is a friend, is willing to falsify some documents to make it appear that I should be given priority over other patients who have been waiting longer than me and whose medical condition is more desperate. I feel deeply grateful to her for this favor, which I know to be criminal; indeed, I am quite explicitly grateful that she was willing to commit this criminal injustice for me. (Roberts 2013, 124) It is understandable that the transplant recipient experiences something akin to gratitude; what the physician did for him is life-altering. Receiving a kidney for transplantation is clearly a benefit and the physician no doubt motivated by benevolence for her patient. But her action is not overall morally meritorious; indeed, her action is wrong. Imagine if she had taken a living person hostage and extracted a kidney from him to transplant into her patient. Roberts says that this "gratitude is immoral because it ascribes overall goodness to an act that is bad" (Roberts 2013, 124). This is correct because gratitude is not triggered merely by beneficence, but rather by morally meritorious beneficence.

Gratitude to those who harm us
Among philosophers, Patrick Fitzgerald changed the terms of the debate when he argued that gratitude is called for in "two anomalous cases" (Fitzgerald 1998). One of these cases is gratitude to those who harm us. He developed his argument in the context of Buddhist teaching. Recently Nicolas Bommarito has contributed to the discussion, defending what he calls "Buddhist gratitude" (Bommarito 2018: 156).

The rationale
Why in the world should a person be grateful to someone who harms her? The Dalai Lama is cited as a paragon of this position. He often tells his audiences that he is grateful to the Chinese. Though the Chinese harmed him, they gave him "the opportunity to practice love for his enemies" and "gave him training in patience and helped his development as a person" (Fitzgerald 1998: 124). Gratitude to those who harm us provides an opportunity to be compassionate, can rid us of destructive emotions like anger, can bring about other benefits, prevent harm, and repair communal ties (127,132,137). But gratitude toward those who harm us is not merely thankfulness for an opportunity to practice virtue. "Gratitude towards those who intend to harm us manifests a concern to care for others, even under the most difficult circumstances, even when they exhibit ill will toward us" (Bommarito 2018: 158). Those who perpetrate harms do not benefit us, but rather provide opportunities, albeit unintentionally (Fitzgerald 1998: 152). One who takes advantage of these opportunities can become a better person, and that is something for which to be thankful. It is natural to think that this is a case of propositional gratitude -being grateful that there are opportunities for moral improvement -but Fitzgerald insists that the positive values mentioned can obtain only if the victim is grateful to those who harmed her (147)(148). What is most distinctive about this account is that "Fitzgerald severs completely the link between the goodwill connecting intention and gratitude" (Gulliford et al. 2013, 304).
There are both conceptual and moral reasons to doubt that gratitude toward those who intend to harm us is an apt case of targeted gratitude.

Conceptual reasons
Philosophers from Seneca to contemporary thinkers have highlighted several features of the moral requirement of gratitude. These include that there is a benefit provided by a benefactor to a beneficiary, that the benefactor acted from benevolence or goodwill and thereby demonstrated concern for the beneficiary, that the beneficiary acknowledge and appreciate the benefactor's morally meritorious conduct, and that the beneficiary is prepared to make a return to the benefactor if a suitable occasion for doing so arises (Seneca 1995/1st century AD;Smith 1982Smith [1790Sidgwick 1981Sidgwick [1907Berger 1975, 299-300;Card 1988, 121-123;McConnell 1993, Chapter II;Roberts 2004, 64;Manela 2016a, 283-284). That the benefactor provides the benefit intentionally and from concern for the beneficiary is part of what makes her conduct morally meritorious. Any third party can recognize this and praise the benefactor, but the beneficiary has a special relationship with her. "The credit we [beneficiaries] give acknowledges that their [benefactors'] good will was of value to us" (Card 1988, 119). And the beneficiary's full gratitude "entails a desire to make a return to the benefactor" (Manela 2016a, 283;Manela 2019, 298). These various accounts are advanced to tease out features that constitute the concept of targeted gratitude. Some psychologists too have understood gratitude to involve these elements. Tesser and colleagues have determined that gratitude is a response to benefits that have been provided intentionally and from altruistic motives (Tesser et al. 1968;see also, McCullough et al. 2001, 254-255). And more recently several psychologists have argued that gratitude involves a recipient's thoughts and perceptions that include acknowledging the goodness of the gift, acknowledging the goodness of the giver, and perceiving the gift as gratuitous (Emmons 2007, 37-38;Watkins 2014, Chapter 3).
Peter Strawson's well-known account of reactive attitudes provides a second source for analyzing the concept of (targeted) gratitude (Strawson 1982(Strawson /1962). Reactive attitudes are reactions people have when they are involved in various transactions with each other. Strawson claims that what he has to say "consists largely of commonplaces" (62), by which he means that these ideas are held by ordinary people and embedded in ordinary language. Two reactive attitudes that Strawson highlights are gratitude and resentment. It matters to us whether the actions of other people "reflect attitudes towards us of goodwill, affection, or esteem on the one hand or contempt, indifference, or malevolence on the other" (63). I would not ordinarily feel resentment if another person trod "on my hand accidentally," but I would feel resentment if he trod on it "in contemptuous disregard of my existence or with a malevolent wish to injure me" (63). Analogously, "[i]f someone's actions help me to some benefit I desire, then I am benefited in any case; but if he intended them so to benefit me because of his general goodwill towards me, I shall reasonably feel gratitude which I should not feel at all if the benefit was an accidental consequence, unintended or even regretted by him, of some plan of action with a different aim" (63). Strawson later makes clear just how commonplace he thinks that these reactive attitudes are. "What I have called the participant reactive attitudes are essentially natural human reactions to good or ill will or indifference of others towards us, as displayed in their attitudes and actions" (67). If these attitudes are that natural, it is not unreasonable to think that they may have an evolutionary basis (Gulliford et al. 2013, 294).
The skeptic might think that Strawson is engaged in armchair psychology for which he has little empirical evidence. This brings us to the third source of support for the conceptual claims, some actual empirical work of psychologists. Joshua Knobe (2003) conducted two simple experiments designed to determine whether people assign intentionality to side effects that are either harmful or beneficial when they are not part of the agent's plan. In one of the studies, a company official went to the chairman of the board with a plan for a new program, which he said would increase profits but harm the environment. The chairman of the board said that he did not care at all about harming the environment as long as the program produced significant profits. This was the "harm condition." In the "help condition," the chairman was told that the new program would increase profits and help the environment. Here too the chairman cared only about increasing profits, and was indifferent about the impact on the environment. In the harm condition, subjects were asked whether the chairman intentionally harmed the environment and how much blame he deserved. In the help condition, they were asked how much praise the chairman deserved and whether the chairman intentionally helped the environment. In the harm condition, 82% of subjects said that the agent did bring about the side effect intentionally, and in the help condition 77% of subjects said that the agent did not bring about the desirable side effects intentionally (Knobe 2003, 191-192). A second experiment produced similar results (192)(193). The two experiments were combined to assess the blame and praise ratings. "Overall, subjects said that the agent deserved a lot of blame in the harm condition (M = 4.8) but very little praise in the help condition (M = 1.4) .... and the total amount of praise or blame that subjects offered was correlated with their judgements about whether or not the side effect was brought about intentionally" (193). From these studies Knobe concludes that "there seems to be an asymmetry whereby people are considerably more willing to blame the agent for bad side effects than to praise the agent for good side effects" (193). Knobe and colleagues did analogous studies with preschool children and found that a preponderance of four-year-old and five-year-old children exhibit a similar "sideeffect effect." They say that a foreseen but bad side effect is brought about "on purpose," but a foreseen morally good side effect is not brought about on purpose (Leslie et al. 2006).
How the intentions of others impact people's helping behavior has even been studied in younger children, and this provides a fourth source of support for the conceptual claim. Amrisha Vaish and colleagues conducted two studies with threeyear-old subjects. They watched an adult either harming or helping another adult (confederates of the researchers). The children subsequently helped the harmful confederate less than they helped a third, neutral adult or the helping adult. The second study had an adult actor trying but failing to harm or to help another adult. Again, the children helped the actor who tried to harm another less than they helped a neutral actor or the helping actor. The authors conclude, "The present studies demonstrate that young children's prosocial behavior is mediated by others' moral behavior" (Vaish et al. 2010(Vaish et al. , 1667. Kristen Dunfield and Valerie Kuhlmeier conducted similar studies with an even younger group of children, twenty-one-monthold infants. In one of the experiments, one actress tried to give a toy to the infant subject but was unable to do so because it accidentally rolled away; the other actress was unwilling to give the toy to the child, pulling it back when the child reached out. Subsequently when these toddlers had their own toys which they could share, they strongly preferred the unable actress over the unwilling one, thus favoring someone who intended to help them. In another experiment, both actresses succeeded in get-ting the toy to the infant, but one deliberately handed it to the subject and the other pretended to drop it in a way that it slid to the subject. When these infants later had their own toys which they could share, they strongly preferred the helping actress whose intentions were unambiguous; in short, they preferentially helped those who had previously willingly given them toys (Dunfield & Kuhlmeier 2010). As Paul Bloom says, "[I]f children are really helping with the interests of someone else in mind, then they should be choosy about whom they help" (Bloom 2013, 52). And research suggests that they are choosy. (For a discussion of related research, see Haidt 2012, 74-75 and 273-274.) It would be nice to have some evidence about how laypeople understand the concept of gratitude and what factors they believe would most impact their experience of gratitude. Fortunately Liz Gulliford and Blaire Morgan have provided us with some such evidence (Gulliford & Morgan 2016), and this is my fifth source for the conceptual claim. This work involves multiple approaches, including prototype analysis, vignette questionnaires, and gratitude stories. I will mention just one small part of this work, that which involved vignette questionnaires. These were designed to explore how various factors influenced laypeople's understanding of gratitude, including cost to the benefactor, value of the benefit, presence of an ulterior motive (other than benevolence) and presence of malicious intention on the part of the benefactor (203). If P2 intended to harm P1, then clearly P2's motive was not benevolence. Indeed, P2's intention is malicious. What Gulliford and Morgan found was that the vast majority of respondents indicated that a malicious intention would undermine their experience of gratitude, though interestingly 12% of the sample "stated that they would be grateful even when the benefit was given with the intention of embarrassing or harming them"(206). The presence of an ulterior motive also significantly lessens the reported experience of gratitude (205-206; see also , Tsang 2006). From these data, we might reasonably speculate that laypeople would report no gratitude at all if the person with whom they were interacting intended only to harm them and not benefit them at all. I conclude that the concept of targeted gratitude requires that benefactors act intentionally from altruistic motives; this precludes the possibility of P1 being aptly grateful to P2 when P2 intended to harm but not to benefit P1. But Fitzgerald (and perhaps Bommarito) can concede this part of the argument. For the point, after all, is reformist, to provide an alternative moral outlook to the view dominant among philosophers. The empirical work cited in the last three reasons, however, suggests that much of the standard philosophical analysis is in line with how laypersons think. But Fitzgerald wants to reform those views too. "So if the contemporary philosophical discussion of gratitude accurately reflects common moral beliefs about when we should be grateful, then those beliefs are wrong" (Fitzgerald 1998, 137). "If gratitude is not owed but can nonetheless cause great benefit or prevent great harm or repair communal ties, then the agent has good moral reason to be grateful" (137). But there are compelling moral reasons to reject this alternative.

Moral reasons
The first thing to note is that the reasons given for being grateful to someone who harms us are instrumental (Gulliford et al. 2013, 312); gratitude in such cases is said to promote other important goods. As we have seen, among these goods are the opportunity to practice and develop such virtues as patience and compassion, a means of ridding oneself of disturbing emotions, especially anger, and a way of promoting peace and communal ties. In some moral traditions, however, especially that of Aristotle, moral virtues and the conduct they lead to have intrinsic value. This opens the door to an objection. "From an Aristotelian perspective, the problem with the instrumentalist routes is that they make the moral trait in question essentially replaceable. If you can find some positive trait that is more conducive to pro-social ends than gratitude, you can substitute gratitude with the other trait" (Kristjansson 2018, 56). If gratitude can be replaced by some other emotion or norm that has greater instrumental value in the situation, "that seems to fly in the face of prevalent moral intuitions about gratitude being an indispensable part of an intrinsically valuable moral life" (Kristjansson 2018, 171). Fitzgerald concedes that gratitude is sometimes replaceable (1998,140). He discusses a case in which a daughter is trying to get her father to let go of anger and resentment against a prior abuser. She suggests that he be grateful to the abuser for giving him an opportunity to deepen his faith. But Fitzgerald notes parenthetically that forgiveness may be a more effective means to the end. And in discussing Faye, a woman who had been systematically abused by her father, Fitzgerald acknowledges (143-144) that there might be several ways to deal with her anger. But if "Faye found that she could not forget, excuse, or forgive", then she might decide "to cultivate gratitude" (144). I find it difficult to fathom that someone could not forgive another but could cultivate gratitude toward him. It is worth noting that something can have both instrumental and intrinsic value. But as the quote in the previous paragraph shows, Fitzgerald asserts that even if gratitude is not owed, agents have reasons to be grateful if doing so brings about great benefits or prevents great harms. And in such cases, gratitude has only instrumental value.
If P1 is not obligated to be grateful to P2 for harming her, but has good moral reasons to do so (Fitzgerald 1998, 138), then how should P1's gratitude be manifested? She will not engage in such harming conduct herself, nor will she recommend such conduct to others. "One is not, for example, advised to go around hurting others to give them a chance to practice responding [with care and compassion] in this way" (Bommarito 2018, 158). Moreover, if P1 is to maintain her self-respect, she will recognize, oppose, stand up to, and criticize the unjust actions of the perpetrators (Fitzgerald 1998, 142). So why should someone display gratitude toward a person who has harmed her? "Gratitude is warranted when an agent accurately notices the direct benefits or the opportunities which another has created for him, and develops appreciation, goodwill, and the disposition to act on that appreciation and goodwill because of those benefits or opportunities" (Fitzgerald 1998, 146). In the standard case of targeted gratitude, the benefit has been provided intentionally and from benevolence, and the appreciation, goodwill, and conduct that follow from that recognition are ways of crediting the benefactor and acknowledging the benefactor's desert. But when a person harms another intentionally, that conduct is criticized not credited. So what is the person harmed supposed to be grateful for?
The answer is opportunities, opportunities that "are given" (126) by the perpetrator, opportunities that the enemy "gives us" (132), opportunities that the perpetrator "has created for" the victim (146). This merges benefits and opportunities in a way that seems plausible only because ordinarily to provide people with opportunities is to provide them with benefits. But to say that the perpetrator of harm "gives" or "creates" opportunities for those harmed is misleading. As Tony Manela puts it, "the agents were not intending their actions under the same description as the actions for which their beneficiaries were grateful" (Manela 2016a, 287). The only thing that the perpetrator did was to create for others a bad situation. The victims admirably turned this situation into an opportunity for moral growth. The perpetrator no more deserves gratitude than does a passerby who unknowingly frightens a mugger, thereby sparing you harm, or a taxi driver whose lateness caused you to miss a plane that subsequently crashes (Simmons 1979, 170-171).
So why not say that propositional gratitude is what is apt in this case? Fitzgerald insists that the victims have moral reasons to be grateful to their assailants because only gratitude to the assailants will rid the victims of their toxic anger (148). But such a claim strains credulity. Forgiveness, compassion, or following the counsel to love thy enemy as thyself (Matthew 5: 43-45) surely are effective antidotes to anger that can replace targeted gratitude. Gratitude's intrinsic value is a beneficiary recognizing and appreciating the benefactor's morally meritorious conduct and being willing to make a return to the benefactor if a suitable occasion for doing so arises (Smith 1982(Smith [1790Price 1974Price [1787Sidgwick 1981Sidgwick [1907, 260-261). The beneficiary has a special reason to prize, not merely praise, the benefactor. This norm is essential to a morally good life and so cannot be replaced.
There are other moral reasons for not accepting this expansionist view of targeted gratitude. Fitzgerald assumes that anger is vice (1998,132) and that anger is morally problematic (148). But a more plausible view is that anger is sometimes fitting and can be justified (Aristotle 1985/4th century BC 1125b 26 -1126a 3). As Robert Roberts states, "Our initial intuitions may tell us that anger is a bad emotion, but a little further reflection may convince us that we would not want to rule it out of human life if we could (we can't). Anger is sometimes on the side of justice and the noble heart. Likewise, gratitude may sometimes be fawning or misconceived" (Roberts 2004, 59). Some psychologists also hold that gratitude and anger are conflicting emotions. Robert Emmons writes that "[g]ratitude drives out toxic emotions of resentment, anger, and envy..." (Emmons 2007, 66). And later he says, "The Buddha said that 'Hatred cannot coexist with loving-kindness.' You cannot be grateful and resentful at the same time, or forgiving and vengeful" (74). But as Liz Jackson points out (Jackson 2016, 286-287), Martin Luther King Jr. could at the same time be angry about white supremacist laws and practices and appreciate some of his own good fortunes and that some things were getting better. As Roberts puts it, "It is possible for a person to resent some things and be grateful for others; the wellfunctioning person will be in just this condition" (Roberts 2004, 69).
Another reason for not accepting this expansionist view of targeted gratitude is that it is at odds with our moral experience. Consider two cases. In the anomalous case, P1 intentionally harms P2 and in so doing manifests no concern for P2's wellbeing. In the standard case, P1 intentionally provides a benefit to P2 and in so doing manifests concern for P2's well-being. P2's targeted gratitude to P1 in the standard case is a response to moral merit; it is a kind of prizing that P2 can most appropriately provide. Even if, following Fitzgerald, P2 is grateful to P1 in the anomalous case, that gratitude is not a response to moral merit; it is qualitatively different from the gratitude in the standard case. Fitzgerald acknowledges that the critic might say that there are "two very different sentiments" (1998,149) in the anomalous and standard cases, but thinks that the objection has no purchase. He asserts, "There is no unique difference in sentiment that we can use to distinguish my anomalous cases from standard cases" (149). He goes on to suggest that a beneficiary's experience of gratitude will vary greatly depending on his relationship with and fondness of the benefactor. He may like her, or he may not like her; he may agree with her political beliefs, or he may disagree. Many factors may influence how he experiences gratitude to her. So the differences in experience between the anomalous cases and the standard cases are of no consequence. "We can have gratitude in all of these cases because it is possible to recognize the value or significance of a person or an act and the role that the person or act plays in our life" (150). But there is a unique difference between the response in the standard case and the anomalous case. In the standard case the beneficiary is responding to an act that manifested concern for his well-being; in the anomalous case, not only was no concern manifested, but hatred or dislike was present. There is intrinsic moral value in a beneficiary's acknowledgement and appreciation of a benefactor's morally meritorious conduct and that is why there is a qualitative difference between the standard case and the anomalous case.
An additional reason for resisting this expansionist view of targeted gratitude is the fear that it may encourage tolerating oppression and injustice (Jackson 2016, 283-286). Being grateful to those who harm us may encourage servility. Fitzgerald thinks that this need not be so. He points out that the Dalai Lama and others maintained self-respect even while being grateful to those who harmed them. "Their feelings of gratitude did not stop them from opposing, standing up to, and criticizing the words and actions of those who persecuted them" (142). There is something a bit odd here. It seems that the Dalai Lama is thanking the Chinese for creating an opportunity to cultivate virtue and simultaneously saying that such conduct is unjust and should cease. There are some strains of thought in the psychological literature that makes Jackson's worries especially important. Because of the emphasis on the positivity of gratitude, psychologists often say that any negative emotions or attitudes cannot coexist with gratitude. Thus, Emmons writes, "The tendency to blame others can be a strong resistance against gratitude" (Emmons 2007, 137). And Watkins says, "The grateful person does not feel that life has been unfair, that they have not received their 'just desserts', and that they are entitled to more benefits than they have received in life" (Watkins 2014, 76). This does not ring true. When, in 1955, Rosa Parks defied Alabama's segregation laws by refusing to sit in the back of the bus, she surely believed that she and others were being treated unjustly. But this need not preclude her being grateful to her friends and family for their love and support (McConnell 2016, 20-21). And slaves during the nineteenth century in the United States had every reason to believe that they were being treated unjustly and denied basic human rights, yet they could express and exhibit gratitude to those who treated them with respect and kindness (McConnell 2017).
There are, then, conceptual reasons for resisting the expansionist account of targeted gratitude. And though some of the proponents of the expansionist account would concede this because they are offering an alternative moral perspective, one that they believe is superior to the more common view, there are multiple reasons for denying that this perspective is morally preferable.

All kinds of grateful persons
If some support being grateful to those who harm us, it seems to be a short step to advocate being grateful for just about anything, or, indeed, everything. In a sermon preached on November 20, 1904, Albert Schweitzer said, "The greatest thing is to give thanks for everything. He who has learned this knows what it means to live. He has penetrated the whole mystery of life: giving thanks for everything" (Schweitzer 1969, 41). Emmons (2007, 98) points out approvingly that gratitude is one of the central motifs in the New Testament writings of Paul. Paul urged Christians to "give thanks in all circumstances" (1 Thess. 5:18) and to "give thanks to God the Father for everything" . Such gratitude is indiscriminate; unless everything one encounters is good and is provided by an agent acting altruistically, this sort of gratitude has ceased to make distinctions. We need to take a step back.

Trait gratitude
Psychologists and philosophers have both been careful to distinguish between episodic/state gratitude, on the one hand, and dispositional/trait gratitude on the other (Roberts 2004, 59-60;Emmons 2007, 34-35, 180;Gulliford et al. 2013, 295-296;Watkins 2014, 21-26). In the episodic sense, gratitude is experienced in a particular case by a person, such as when someone helps her in a difficult situation. By contrast, one who possesses the trait of gratitude is disposed to experience gratitude in all situations of a certain type, such as when being helped by another in a time of need. In the psychological literature, the two best known instruments for measuring gratitude are GQ6 (McCullough et al. 2002) and GRAT (Watkins et al. 2003), and each of these instruments is designed to measure the trait of gratitude, to identify the extent to which an individual is a grateful person.
There have been some mild criticisms of these instruments and I think that they reveal something important. Lambert and colleagues remark that "several of Mc-Cullough et al.'s scale items can clearly be perceived as not including such an interpersonal transfer of a benefit but instead as measuring more generalized gratitude" (Lambert et al. 2009(Lambert et al. , 1205. They go on to assert that "much of Watkins et al.'s GRAT scale does not measure such interpersonal exchanges" (1205). More recently, Morgan and colleagues make similar points: "The majority of items in existing gratitude measures aim to assess grateful emotions only. Most notable is the GQ6, where all 6 items arguably assess feelings of gratitude." Continuing, they say, "A second component of gratitude is behavior: for instance, expressions of thanks or recognition of others' beneficence. Yet this element of grateful experience is missing from GQ6 and barely features in GRAT" (Morgan et al. 2017, 180;see also, Kristjansson 2018, 54). Mendonca and Palhares (2018) make a similar point, arguing that GQ6 does not in any way measure whether people who have been helped by another feel any sense of obligation to reciprocate (94; see also, Tudge and Freitas 2018: 4-5). These criticisms are important, but there is a deeper conceptual problem that underlies these difficulties. I shall argue that there are several notions of 'grateful person' at work in the psychological literature, that the best known instruments fail to measure the most morally significant notion, and as a result more work is needed to show what the relationship is between the moral virtue of gratitude and subjective well-being.

Two kinds of grateful persons
A common portrayal of grateful persons in the psychological literature says, "Grateful thinking fosters the savoring of positive life experiences and situations, so that people can extract the maximum possible satisfaction and enjoyment from their circumstances" (Emmons 2007, 35). Emmons later says, "A dispositionally grateful person will tend to see what is good in situations and to notice less what is bad" (182). In this same vein, Philip Watkins says that "the affective trait of gratitude may make people more likely to notice positive events" (Watkins 2014, 104). And O'Brien and colleagues point out that some psychologists view gratitude as an emotion, one just as likely to be felt when viewing a beautiful sunset as when another provides help (O'Brien et al. 2018, 180).
But the psychological literature also states clearly and frequently that a grateful person is someone who recognizes that a benefit has been provided, that it was good, that the giver of the gift is good, and that the gift was provided gratuitously (Emmons 2007, 37-38 and 182;Watkins 2014, 42-49). O'Brien and colleagues say that, understood as a virtue, gratitude involves noticing when others help one and responding appropriately (O'Brien et al. 2018, 180-181). So the term 'grateful person' is not used univocally. (Perhaps some of these authors are what Tony Manela [2016a, 292] calls "inclusivists," persons who hold that a unified account of the various types of gratitude is possible. Like Manela, I believe that the concepts are so different that a unified account is not possible. But in any case, inclusivists will have to show what the unified account is.) Earlier I distinguished between propositional (dyadic) gratitude, being grateful that a certain state of affairs exists (for example, grateful that the weather is good for our family picnic, that I got a job teaching philosophy, that my mother's cancer is in remission) and targeted (triadic) gratitude, being grateful to someone for a benefit she has provided me (for example, grateful to Jane for giving me a ride home, grateful to Joan for providing emotional comfort when I was depressed). It is this distinction, I think, that accounts for the different senses of 'grateful person' just described. These two types of gratitude are different in kind, and that suggests that there are two different (though possibly related) traits. Corresponding to propositional gratitude, a p-grateful person is someone who is disposed to see what is good in situations, is adept at identifying, appreciating, and utilizing those good things, and who generally thinks positively. Corresponding to targeted gratitude, a t-grateful person is disposed to recognize and appreciate when someone has engaged in morally meritorious conduct from which he has benefited and is disposed to respond with the proper degree of grateful behavior (Roberts 2004, 60;Morgan et al. 2015, 101;Manela 2019, 303-304).

Why this distinction matters
The reason that this distinction matters is that p-gratefulness and t-gratefulness are different traits and some of the things true of one need not be true of the other. There is good reason to think that p-gratefulness will be associated with agreeableness, optimism, and positive feelings. But t-gratefulness need not always be associated with these emotions; it might depend in part on how often a person is treated with kindness by others. It is often claimed that if someone sees himself as a victim or believes that life has been unfair, this impedes the development of gratitude (Emmons 2007, 180;Watkins 2014, 76); but while this may impede the development of p-gratefulness, it need not thwart t-gratefulness. Indeed, t-gratefulness may be enhanced if someone who is routinely treated unfairly is on occasions treated with kindness by another particular person. When Emmons says, "Gratitude is the realization that we have everything that we need, at the moment" (137), this reflects pgratefulness but not t-gratefulness. Another reason that the distinction is important is that researchers must be clear what concept they are investigating if their results are to be reliable (O'Brien et al. 2018, 181). One other point is relevant here. Philip Watkins makes this astute observation: "It is important to distinguish between gratitude interventions that are designed to enhance well-being, and interventions that are designed to enhance grateful responding (the disposition of gratitude)" (Watkins 2014, 231). Gratitude interventions include counting one's blessings, keeping a gratitude journal, making gratitude visits, engaging in gratitude reappraisal, among others (Emmons 2007, Chapter 2;Watkins 2014, Chapter 13). Notice that Watkins' remark can be paraphrased to ask, "Is the intervention designed to promote p-gratefulness, t-gratefulness, or both?" In a therapeutic setting, we might expect the focus to be on p-gratefulness; if the task is moral education, t-gratefulness seems to be the relevant trait. This shows that the traits are distinct and require distinct measures.
Only one of these traits, t-gratefulness, is obviously a moral virtue (O'Brien et al. 2018, 179-181). Its intrinsic value is that its possessors are disposed to recognize, acknowledge, and appreciate that they have benefited from another's morally meritorious conduct, and are disposed to help the benefactor if a suitable occasion arises. Is p-gratefulness a virtue? It seems clear that it is a prudential virtue (Manela 2016a, 289, says that propositional gratitude is a sort of "prudential appreciation."). Being disposed to make the best of a bad situation is clearly in one's interest; any trait that better enables a person to cope with difficulties is advantageous. Moreover, there is overwhelming evidence that p-gratefulness contributes positively to mental and physical health, handling challenging situations, and life satisfaction in general (Emmons 2007 andWatkins 2014). But having a positive outlook on life, being adept at seeing the good in bad situations, and being able to extract opportunities from challenges does not initially seem to be a trait that one would say is a moral virtue. It is not obvious, for example, that someone with a positive outlook on life will also be disposed to help a prior benefactor if a suitable occasion arises.
There are at least two ways that one might connect p-gratefulness to moral virtues. A disposition to improve oneself morally is, I would maintain, a second-order moral virtue, a disposition to improve and better cultivate first-order moral virtues. Persons who are p-grateful may be especially adept at seeing opportunities for moral improvement. The case of the Dalai Lama discussed earlier illustrates this. This does not make p-gratefulness a moral virtue, but rather suggests that it is an enabling virtue, a trait that better enables its possessor to improve morally. There is a second, and perhaps more direct, way that p-gratefulness may be thought of as a moral virtue. If there are good moral reasons to take care of oneself, then there may be moral reasons to cultivate p-gratefulness because it contributes positively to good physical and mental health. But even if this is true, these reasons are different from the moral reasons to cultivate t-gratefulness; for the latter are triggered by recognizing the moral desert of one's benefactors.
Some of the most enthusiastic pronouncements about trait gratitude show why making this distinction is important. We are told, "A person with the disposition to feel grateful has established a worldview that says, in effect, that all of life is a gift, gratuitously given" (Emmons 2007, 187; see also, Watkins 2014, 22, 87-88). But surely there is no reason to think that someone who possesses the moral virtue of t-gratefulness will necessarily believe that all of life is a gift. P-gratefulness involves identifying something as good and enjoying it; it is a paradigmatically agreeable or pleasant feeling (Manela 2016a, 292). T-gratefulness, however, need not always be pleasant. While it is, of course, normally pleasant to receive a benefit from another motivated by benevolence, there can be exceptions. Perhaps one does not want a further relationship with the benefactor (Card 1988, 125), or perhaps the benefactor's sacrifices were so extreme that the beneficiary laments the harm that the benefactor experienced in providing the help (Manela 2016b). Some gratitude enthusiasts have expressed puzzlement about the fact that "when people report that they are feeling grateful, they also tend to report feeling somewhat indebted" (Watkins 2014, 216). This is a "conundrum" (216) because indebtedness is thought to be a negative emotion. This is a puzzle, however, only if we think that only positive emotions are associated with gratitude. And while that may be the case with p-gratitude, it is not so with t-gratitude. Emmons notices this when he writes, "To be grateful means to allow oneself to be placed in the position of a recipient -to feel indebted and aware of one's dependence on others" (Emmons 2007, 29).
Because p-gratefulness and t-gratefulness are different character traits, it would be desirable to have separate measurements for them. If the observations noted earlier (Lambert et al. 2009(Lambert et al. , 1205Morgan et al. 2017, 180;Kristjansson 2018, 54) are correct, it seems that GQ6 and GRAT primarily measure p-gratefulness. Morgan and colleagues (2017) have developed an instrument, called the Multi-Component Gratitude Measure (MCGM), that tries to go beyond this. It is a method designed to assess four distinct dimensions of the virtue of gratitude: conceptions of gratitude, grateful emotions, attitudes toward gratitude, and gratitude-related behavior. Because these dimensions are distinct, MCGM does not provide "one simple 'gratitude score'" (Morgan et al. 2017, 188). The dimensions of "attitudes toward gratitude" and "gratitude-related behavior" are especially relevant to t-gratefulness. As such, MCGM goes beyond previous measures and tells us what subjects believe about when gratitude is due and what constitutes grateful behavior; it is not open to the objection of Lambert et al. (2009Lambert et al. ( , 1205) that it does not take into account the relevance of interpersonal transfer of benefits. It is important to note, however, that MCGM does not purport to tell us whether people are in fact t-grateful (I am skeptical whether we can determine experimentally whether a person genuinely exhibits t-gratefulness. See McConnell 2018, 199-200.). Nor does it tell us whether moral educators should concern themselves with trying to make children grateful; on this view I share the skepticism of Morgan and her colleagues (Morgan et al. 2015, 104).

A challenge to the distinction
One way to defend what I am calling expansionist views of gratitude is to deny the importance of the distinction between p-gratefulness and t-gratefulness. Some may suspect that p-gratefulness involves religious or metaphysical assumptions (as was suggested to me by Marie-Luise Raters). This may prompt critics to argue that all of the cases in which we experience gratitude are cases of t-gratefulness. Emmons describes this possibility.
When the blessings that we have cannot be attributed to human benevolence, attributions to God's goodness become all the more likely. Therefore, people are more likely to sense a divine hand in cherished experiences that cannot easily be attributed to human effort -the birth of a child, a miraculous recovery from illness, the restoration of an estranged relationship -for which gratitude to God is the apt response. (Emmons 2007, 108) Others have also maintained that when people experience gratitude, they are implicitly attributing benevolent intentions to some giver, such as God (Watkins 2014, 46, 93). This approach reduces all cases of gratitude to t-gratitude. There are two weaknesses in this view. First, it is not true that every person who expresses gratitude that a certain state of affairs exists ascribes this existence to the agency of another being. At least some are simply saying that they are glad, that they feel fortunate, or that they appreciate that a beneficial state of affairs exists. Second, even among those who sometimes ascribe good fortune to a divine being, many do not always make such an ascription. Some things for which we might be grateful seem too trivial to have been the result of divine intervention. If I arrive at my bus stop two minutes late but the bus driver is four minutes behind schedule, I may say that I am grateful that this state of affairs allowed me to get to work on time. But would anyone really attribute the bus driver being late to divine intervention? Some recent work in neuroscience lends support to (though certainly does not prove) the legitimacy of the distinction. Christina Karns studies brain activity that is associated with moral emotions, including gratitude. She employs a narrative approach. Research participants read scenarios that are designed to evoke gratitude and then brain activity is measured (Karns 2019, 206). In one study, narrative vignettes were presented that distinguished merely receiving a benefit (for example, finding the last empty parking spot in a crowded lot) from having a benefit bestowed by another (for example, another person waving you in to the only remaining spot). Karns's research, as well as that of others, demonstrates that different regions of the brain show differential responses depending on whether agency was perceived to be involved in the provision of the benefit (Karns 2019, 206-210). If human beings process receiving benefits from an agent differently from merely encountering a benefit in the world, perhaps there are evolutionary advantages in distinguishing p-gratefulness from t-gratedfulness. This comports with Marc Hauser's broader arguments that it is an essential feature of moral creatures that they understand the intentional-accidental distinction (Hauser 2006, 49-52).

Conclusion
Gratitude to someone who has harmed us is inapt because targeted gratitude is fitting only when a benefit has been provided by a morally admirable act motivated by benevolence. Instances of p-gratitude are not per se inapt; but when they are presented along with instances of t-gratitude as instances of the same kind of gratitude, the moral distinctness of gratitude is diluted. And when enthusiasts endorse what seems to be indiscriminate gratitude, the very essence of the concept has been lost. Robert Roberts, who does distinguish these two senses of gratefulness, complains, "It would be better to use the other expression that I have suggested [gladness], and to reserve gratitude for the three-term construal" (Roberts 2004, 63). This is not mere caviling. T-gratefulness is a moral virtue, while p-gratefulness is a prudential trait. It would be less confusing to have separate terms. But ordinary language being what it is, we have to endure laborious distinctions.